TY - JOUR
T1 - Diagnostic utility of a likert scale versus qualitative descriptors and length of capsular contact for determining extraprostatic tumor extension at multiparametric prostate MRI
AU - Costa, Daniel N
AU - Passoni, Niccolo M.
AU - Leyendecker, John
AU - Diaz De Leon III, Alberto
AU - Lotan, Yair
AU - Roehrborn, Claus
AU - Otero-Muinelo, Susana
AU - Grewal, Harpreet
AU - Xi, Yin
AU - Francis, Franto
AU - Rofsky, Neil M
AU - Pedrosa, Ivan
N1 - Funding Information:
Supported by grant UL1TR001105 from the National Center for Advancing Translational Sciences of the National Institutes of Health.
Publisher Copyright:
© American Roentgen Ray Society.
PY - 2018/5
Y1 - 2018/5
N2 - OBJECTIVE. The purpose of this study is to determine the reproducibility and diagnostic performance of a Likert scale in comparison with the European Society of Urogenital Radiology (ESUR) criteria and tumor-pseudocapsule contact length (TCL) for the detection of extraprostatic extension (EPE) at multiparametric MRI. MATERIALS AND METHODS. This was a retrospective review of all men who underwent multiparametric MRI followed by prostatectomy between November 2015 and July 2016. Multiparametric 3-T MRI studies with an endorectal coil were independently reviewed by five readers who assigned the likelihood of EPE using a 1-5 Likert score, ESUR criteria, and TCL (> 10 mm). EPE outcome (absent or present) for the index lesion at whole-mount histopathologic analysis was the standard of reference. Odds ratios (ORs) and areas under the ROC curve (A z ) were used for diagnostic accuracy. The interreader agreement was determined using a weighted kappa coefficient. A p < 0.05 was considered significant. RESULTS. Eighty men met the eligibility criteria. At univariate analysis, the Likert score showed the strongest association (OR, 1.8) with EPE, followed by prostate-specific antigen level (OR, 1.7), ESUR score (OR, 1.6), and index lesion size (OR, 1.2). At multivariable analysis, higher Likert score (OR, 1.8) and prostate-specific antigen level (OR, 1.6-1.7) were independent predictors of EPE. The A z value for Likert scores was statistically significantly higher (0.79) than that for TCL (0.74; p < 0.01), but not statistically significantly higher than the value for ESUR scores (0.77; p = 0.17). Interreader agreement with Likert (κ = 0.52) and ESUR scores (κ = 0.55) was moderate and slightly superior to that for TCL (κ = 0.43). Except for TCL among inexperienced readers (κ = 0.34), reader experience did not affect interreader agreement. CONCLUSION. A Likert score conveying the degree of suspicion at multiparametric MRI is a stronger predictor of EPE than is either ESUR score or TCL and may facilitate informed decision making, patient counseling, and treatment planning.
AB - OBJECTIVE. The purpose of this study is to determine the reproducibility and diagnostic performance of a Likert scale in comparison with the European Society of Urogenital Radiology (ESUR) criteria and tumor-pseudocapsule contact length (TCL) for the detection of extraprostatic extension (EPE) at multiparametric MRI. MATERIALS AND METHODS. This was a retrospective review of all men who underwent multiparametric MRI followed by prostatectomy between November 2015 and July 2016. Multiparametric 3-T MRI studies with an endorectal coil were independently reviewed by five readers who assigned the likelihood of EPE using a 1-5 Likert score, ESUR criteria, and TCL (> 10 mm). EPE outcome (absent or present) for the index lesion at whole-mount histopathologic analysis was the standard of reference. Odds ratios (ORs) and areas under the ROC curve (A z ) were used for diagnostic accuracy. The interreader agreement was determined using a weighted kappa coefficient. A p < 0.05 was considered significant. RESULTS. Eighty men met the eligibility criteria. At univariate analysis, the Likert score showed the strongest association (OR, 1.8) with EPE, followed by prostate-specific antigen level (OR, 1.7), ESUR score (OR, 1.6), and index lesion size (OR, 1.2). At multivariable analysis, higher Likert score (OR, 1.8) and prostate-specific antigen level (OR, 1.6-1.7) were independent predictors of EPE. The A z value for Likert scores was statistically significantly higher (0.79) than that for TCL (0.74; p < 0.01), but not statistically significantly higher than the value for ESUR scores (0.77; p = 0.17). Interreader agreement with Likert (κ = 0.52) and ESUR scores (κ = 0.55) was moderate and slightly superior to that for TCL (κ = 0.43). Except for TCL among inexperienced readers (κ = 0.34), reader experience did not affect interreader agreement. CONCLUSION. A Likert score conveying the degree of suspicion at multiparametric MRI is a stronger predictor of EPE than is either ESUR score or TCL and may facilitate informed decision making, patient counseling, and treatment planning.
KW - Imaging-pathology correlation
KW - MRI
KW - Prostate cancer
KW - Risk stratification
KW - Staging
UR - http://www.scopus.com/inward/record.url?scp=85046142119&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85046142119&partnerID=8YFLogxK
U2 - 10.2214/AJR.17.18849
DO - 10.2214/AJR.17.18849
M3 - Article
C2 - 29489410
AN - SCOPUS:85046142119
SN - 0361-803X
VL - 210
SP - 1066
EP - 1072
JO - American Journal of Roentgenology
JF - American Journal of Roentgenology
IS - 5
ER -